Global Investigation of an Engineered Nitrogen-Fixing Escherichia Coli

www.nature.com/scientificreports

OPEN Global investigation of an engineered nitrogen-fxing Escherichia coli strain reveals Received: 18 December 2017 Accepted: 6 July 2018 regulatory coupling between host Published: xx xx xxxx and heterologous nitrogen-fxation genes Zhimin Yang1,2, Yunlei Han2, Yao Ma2, Qinghua Chen2, Yuhua Zhan2, Wei Lu2, Li Cai1, Mingsheng Hou1, Sanfeng Chen3, Yongliang Yan2 & Min Lin2

Transfer of nitrogen fxation (nif) genes from diazotrophs to amenable heterologous hosts is of increasing interest to genetically engineer nitrogen fxation. However, how the non-diazotrophic host maximizes opportunities to fne-tune the acquired capacity for nitrogen fxation has not been fully explored. In this study, a global investigation of an engineered nitrogen-fxing Escherichia coli strain EN-01 harboring a heterologous nif island from Pseudomonas stutzeri was performed via transcriptomics and proteomics analyses. A total of 1156 genes and 206 discriminative proteins were found to be signifcantly altered when cells were incubated under nitrogen-fxation conditions. Pathways for regulation, metabolic fux and oxygen protection to nitrogenase were particularly discussed. An NtrC-dependent regulatory coupling between E. coli nitrogen regulation system and nif genes was established. Additionally, pentose phosphate pathway was proposed to serve as the primary route for glucose catabolism and energy supply to nitrogenase. Meanwhile, HPLC analysis indicated that organic acids produced by EN-01 might have negative efects on nitrogenase activity. This study provides a global view of the complex network underlying the acquired nif genes in the recombinant E. coli and also provides clues for the optimization and redesign of robust nitrogen-fxing organisms to improve nitrogenase efciency by overcoming regulatory or metabolic obstacles.

In nature, a variety of genes and islands can be rapidly and frequently horizontally transferred among bacteria, resulting in the acquisition of certain properties such as nitrogen fxation, antimicrobial resistance and patho- genesis, which help bacteria to succeed in altered habitats or new niches1–3. However, newly acquired genes or islands become a burden for bacteria if they are not properly integrated with host regulatory systems. A greater understanding of the physiological alterations that occur in an engineered cell following the insertion of large fragments of foreign DNA, especially from distant species, is needed to pave the way towards the goal of biological engineering. Biological nitrogen fxation is catalyzed, in most cases, by the molybdenum nitrogenase encoded by a highly conserved nifHDK gene cluster. Previous studies have shown that the nif genes encoding active nitrogenase can be transferred to non-nitrogen-fxing prokaryotes to impart the ability to reduce atmospheric nitrogen gas into ammonia as a nitrogen source4–12. From the perspective of synthetic biology, one key goal of studying biological

1Key Laboratory of Plant Pathology of Hubei Province, College of Plant Science and Technology, Huazhong Agricultural University, Wuhan, 430070, China. 2National Key Facility for Crop Gene Resources and Genetic Improvement, Biotechnology Research Institute, Chinese Academy of Agricultural Sciences, Beijing, 100081, China. 3State Key Laboratory of Agrobiotechnology and College of Biological Science, China Agricultural University, Beijing, 100193, China. Correspondence and requests for materials should be addressed to Y.Y. (email: yanyongliang@caas. cn) or M.L. (email: [email protected])

SCieNTiFiC REPOrts | (2018)8:10928 | DOI:10.1038/s41598-018-29204-0 1 www.nature.com/scientificreports/

nitrogen fxation is to facilitate the introduction of this ability into organisms of great importance for human beings, for instance, engineering autonomous nitrogen-fxing cereal crops13–15. Eventually, a successfully engineered N2-fxing non-legume crop may signifcantly cut down the use of chemical fertilizers for a cleaner environment and higher yields16,17. In recent years, the synthesis of nitrogen-fxing systems has become increasingly common due to advances in synthetic biology. Normally, a functional entity or pathway is frst detected in prokaryotes before being transferred to eukaryotes and plants. Because of its well-studied genetic background, Escherichia coli is frequently used as the preferred frst-step research model. Following the pioneering work on nitrogen-fxation engineering in 1970s4,5, several groups have reported successful gene transfer of nif genes to E. coli in the past fve years11,12,18,19. However, these recombinant E. coli stains showed much lower nitrogenase activity compared with the original host11,12,, and the horizontally acquired ability was insufcient to enable diazotrophic growth on nitrogen-free medium, implying the presence of (i) regulatory coupling between the host and heterologous nitrogen-fxation systems, as well as (ii) a regulatory/or metabolic barrier that results in reduced nitrogenase activity in the engineered cells. To date, how the non-diazotrophic host maximizes opportunities to fne-tune the acquired capacity for nitrogen fxation has not yet been fully explored. Pseudomonas stutzeri A1501 is a root-associated bacterium that exhibits an unusual feature, for a Pseudomonas strain, the ability to fx nitrogen20–24. Te P. stutzeri A1501 genome contains a 49-kb nitrogen fxation island (NFI) that comprises the largest group of nif genes identifed to date25. Within this island, a total of 52 nif-related genes are organized into 11 putative NifA-δ54-dependent operons24. nif gene expression in A1501 was revealed to be tightly regulated at both the transcriptional and post-transcriptional levels22,23,26,27. Given its natural integrity and well-studied regulation, the A1501 NFI is a promising model for studying the synthetic biology of nitrogen fxation systems. We previously transferred the entire P. stutzeri A1501 NFI into E. coli and found that the nitrogenase activity of the engineered E. coli strain was dependent on the external ammonium concentration, oxygen tension and temperature12. Similar to previous reports, the nitrogenase activity of recombinant E. coli strain EN-01 was much lower than that of A1501. In the present study, to better understand the global regulatory efect of the host on NFI expression, we monitored the global transcriptional and proteomic profles of recombinant E. coli grown anaerobically under nitrogen-fxation and nitrogen-repression conditions. Furthermore, the metabolic fux shif of E. coli EN-01 under nitrogen-fxation conditions was also determined by HPLC. To the best of our knowledge, this is the frst report of a global investigation into the regulatory cascade of nif genes in an engineered nitrogen-fxing organism. Tese data are particularly useful for providing a more comprehensive understanding of how the E. coli host intervenes in the transcriptional regulation of the “foreign” NFI to support a functional nitrogenase complex. Our results will also help guide eforts to more successfully remodel and optimize similar systems in other species. Results Overview of the E. coli EN-01 transcriptome and proteome under nitrogen-fxation and nitrogen-repression conditions. Te expression of 1156 genes was signifcantly altered (≥2-fold, P ≤ 0.05) under nitrogen-fxation conditions compared with nitrogen-repression conditions, including 789 up-regulated and 367 down-regulated genes. Te up-regulated genes mainly belonged to three major functional categories: nitrogen metabolism, transport or membrane protein and unknown function, while the down-regulated genes were mainly involved in energy synthesis, transport, protein synthesis and regulation. Tese altered genes were further classifed according to the COG functional classifcation system. As shown in Fig. 1, 259 genes, of which 102 were down-regulated and 157 were up-regulated, were involved in bacterial preservation and processing of genetic information (DNA replication, duplication, repair and gene transcription, expression, etc.). A total of 361 genes (31% of the total altered genes) were involved in transport and metabolic pathways. An additional 90 genes were involved in energy synthesis and transformation processes, with 67 genes induced under nitrogen-fxation conditions (Fig. 1A). Moreover, the expression of 151 genes encoding proteins of unknown function was also signifcant altered. Transcriptome analysis has revealed that 255 genes were significantly up-regulated and 294 genes were severely down-regulated in P. stutzeri A1501 under nitrogen-fxation conditions25. Subsequent comparison of the E. coli EN-01 and P. stutzeri A1501 transcriptomes revealed that at least 24 genes in the E. coli EN-01 transcriptome showed a similar expression pattern to that in P. stutzeri A1501 (Supplementary Table S1), including the glnK-amtB operon, the two-component regulatory system ntrBC andthe serine protein kinase-coding gene prkA, implying that the overlapping genes or systems might be essential for NFI expression in E. coli. Proteomic analysis of E. coli EN-01 was also performed under identical conditions. A total of 110 protein spots increased by more than 2-fold under nitrogen-fxation conditions; notably, 55 proteins were only expressed under nitrogen-fxation conditions. In addition, 96 proteins were signifcantly down-regulated under nitrogen-fxation conditions, and the expression of 24 protein spots was abolished on the 2D-PAGE gel (Supplementary Fig. S1). A total of 138 proteins were further identifed by MALDI-TOF mass spectrometry, of which 63 were up-regulated and 75 were down-regulated (Supplementary Table S2). In addition, 132 identified proteins showed similar expression patterns to the transcriptional analysis. Tese proteins included the glutamine synthetase (GS) GlnA, nitrogen regulatory PII protein GlnK, hydroperoxidase II KatE, oligopeptide ABC transporter OppA, glucose-6-phosphate 1-dehydrogenase Zwf and the dihydrolipoamide dehydrogenase LpdA. Taken together, the above data revealed dramatic changes to the global transcriptional regulatory networks and cellular metabolic pathways in the recombinant nitrogen-fxing E. coli EN-01 under nitrogen-fxation conditions. Subsequent analyses focused on the nitrogen regulatory system, energy production and conversion systems, and the oxygen protection-related pathway, which have been reported to be crucial for nitrogen fxation28.

SCieNTiFiC REPOrts | (2018)8:10928 | DOI:10.1038/s41598-018-29204-0 2 www.nature.com/scientificreports/

Figure 1. Overview of expression profling analysis in recombinant nitrogen-fxing E. coli EN-01. (A) Functional categories of nitrogen fxation-induced genes (P ≤ 0.05 and fold change ≥2) in E. coli EN-01. (B) Functional categories of core subset of down-regulated genes (P ≤ 0.05 and fold change ≥2) under nitrogen- fxation conditions. Te percentage of genes in each section is depicted.

Expression patterns of E. coli nitrogen metabolic pathways and regulatory system under nitrogen-fxation conditions. Expression of P. stutzeri NFI genes in recombinant E. coli strain EN-01. EN- 01 exhibits nitrogenase activity under glucose-fermentation and nitrogen-fxation conditions, indicating the heterologous production of a functional nitrogenase complex12. To further test the heterologous expression of nif genes, we determined the induction ratio of 43 genes within the island using quantitative real-time RT-PCR (Table 1). Previously, we found that the A1501 NFI genes are organized into 11 NifA-σ54-dependent operons and are up-regulated under nitrogen-fxation conditions24. Similarly, except for a 7-gene operon (PST1302-1306), all NifA-σ54-dependent operons were strongly induced in E. coli, suggesting that the nif-specifc transcriptional activator NifA is controlled by the E. coli nitrogen regulatory system. Consistent with the RT-PCR results, proteomic analysis also revealed that the NifS, NifD and PST1336 proteins encoded by the NFI genes were up-regulated under nitrogen-fxation conditions (Fig. 2). Why the PST1302-1306 operon (including the nifQ and nifB genes) was not regulated in EN-01 is still unclear. Te nifQ mutant of Klebsiella pneumoniae is defective in nitrogen fxation due to an elevated requirement for molybdenum29. In addition, nifB has been long recognized as crucial

SCieNTiFiC REPOrts | (2018)8:10928 | DOI:10.1038/s41598-018-29204-0 3 www.nature.com/scientificreports/

Up-regulation folds Gene ID Gene name Heterologous expression in E. coli Expression in P. stutzeri A1501a PST1302 PST1302 0.68 ± 0.04 16.83 PST1303 PST1303 0.77 ± 0.12 53.99 PST1304 nifQ 1.03 ± 0.14 46.56 PST1305 PST1305 1.11 ± 0.36 38.67 PST1306 nifB 0.73 ± 0.22 21.46 PST1313 nifA 7.53 ± 0.56 6.95 PST1314 nifL 5.58 ± 0.31 7.68 PST1316 rnfB 6.8 ± 1.06 9.94 PST1317 rnfC 6.59 ± 0.69 2.22 PST1318 rnfD 6.57 ± 0.74 7.67 PST1319 rnfG 5.96 ± 0.87 7.47 PST1320 rnfE 6.02 ± 1 8.80 PST1321 rnfH 6.09 ± 0.38 17.55 PST1322 nifY 8.67 ± 1.25 21.74 PST1325 PST1325 8.99 ± 0.3 9.68 PST1326 nifH 69.51 ± 8.95 94.05 PST1327 nifD 32.36 ± 1.79 54.16 PST1328 nifK 43.88 ± 2.37 38.22 PST1329 nifT 8.21 ± 0.84 7.82 PST1330 nifY 6.24 ± 0.61 8.51 PST1331 PST1331 5.76 ± 0.7 12.55 PST1332 PST1332 5.16 ± 0.47 3.27 PST1333 nifE 68.88 ± 6.88 35.82 PST1334 nifN 70.35 ± 5.08 13.32 PST1335 nifX 70.01 ± 6 37.97 PST1336 PST1336 39.56 ± 5.22 5.06 PST1341 PST1341 14.21 ± 0.82 1.28 PST1342 PST1342 23.42 ± 2.81 3.69 PST1344 PST1344 11.44 ± 0.94 7.16 PST1345 modC 11.59 ± 1.23 1.59 PST1346 modB 9.08 ± 2.51 2.05 PST1347 modA 7.92 ± 0.68 4.12 PST1348 PST1348 6.86 ± 1.02 3.88 PST1350 nifU 103.22 ± 10.52 10.77 PST1351 nifS 51.33 ± 8.98 16.21 PST1352 nifV 41.38 ± 4.1 24.55 PST1353 cysE 17.32 ± 1.85 32.98 PST1354 PST1354 11.3 ± 0.6 11.11 PST1355 nifW 6.8 ± 0.49 26.04 PST1356 nifZ 12.06 ± 1.46 18.17 PST1357 nifM 7.42 ± 1.01 18.78 PST1358 clpX2 12.67 ± 1.56 8.42 PST1359 nifF 16.81 ± 0.69 14.91

Table 1. Heterologous NFI gene expression in E. coli EN-01 under nitrogen-fxation conditions and comparison with the corresponding genes in P. stutzeri A1501. aTranscriptional ratios of P. stutzeri A1501 NFI genes were obtained from DNA microarray experiments25.

for nitrogen fxation because NifB participates in the biosynthesis of the FeMo-co factor30. We therefore postulate that the relatively low nitrogenase activity of EN-01 might be partly attributed to low expression of these factors within the operon.

Regulatory coupling between the E. coli general nitrogen regulatory system and the heterologous P. stutzeri nif island. Regulation of the nitrogen-fxation process in free-living diazotrophs facilitates the stringent control that is necessary to maximize the physiological benefts from diazotrophy28. In P. stutzeri A1501, the nitrogen regulatory cascade comprises the AmtB–GlnK–NtrBC global nitrogen regulation proteins, which sense the nitrogen signal and subsequently control the expression of the nif-specifc regulatory proteins NifLA31. However, the nitrogen regulatory network in E. coli appears to be an expanded version compared to that in Pseudomonas,

SCieNTiFiC REPOrts | (2018)8:10928 | DOI:10.1038/s41598-018-29204-0 4 www.nature.com/scientificreports/

Figure 2. Enhanced expression of nitrogen fxation related proteins in recombinant E. coli EN-01 under nitrogen-fxation conditions. Magnifed regions of 2-D gel images were sliced from Supplementary Fig. S1. Protein spots of interest are indicated with circles with the corresponding protein names pointed out on the lef. Proteins were identifed by MALDITOF-MS analysis.

with one additional PII protein and cascade regulation achieved by a Nac regulatory protein32,33. Acquisition of a nif-specifc regulatory system in the recombinant E. coli raises the question of nitrogen-fxation efciency by these two regulatory systems of diferent evolutionary origins. Changes in the expression of E. coli nitrogen regulatory system genes are summarized in Supplementary Table S3. Of these, the glnLG genes, which are homologs of Pseudomonas ntrBC, were up-regulated by more than 10-fold, the sigma factor rpoN was up-regulated by 3.64-fold, the ammonium transporter amtB was up-regulated by 55.28-fold, and the two PII protein genes glnK and glnB were both up-regulated under nitrogen-fxation conditions. Consistent with the transcriptional change, proteomic data indicated that GlnK and GlnA expression was enhanced (Supplementary Table S2). Based on the observations in the present study and those from previous experiments from a wide range of model systems, we propose a regulatory cascade that controls heterologous nif transcription in E. coli EN-01 (Fig. 3): Te UTase/URase enzyme GlnD (glnD gene was up-regulated 5.27-fold) is activated by low internal glutamine concentrations and further modifes the PII protein GlnB (glnB was up-regulated 2.12-fold) and GlnK (glnK was up-regulated 86.88-fold) to PII-UMP. Subsequently, PII-UMP phosphorylates NtrC. As a global transcriptional regulator, NtrC activates the expression of a series of nitrogen

SCieNTiFiC REPOrts | (2018)8:10928 | DOI:10.1038/s41598-018-29204-0 5 www.nature.com/scientificreports/

Figure 3. Proposed cascade regulation of nif genes in E. coli EN-01 under nitrogen-fxation conditions. A transcriptional regulatory network was constructed from mRNA and protein expression data. Red arrows, induction. Numbers in red indicates the transcriptional up-regulation ratio of protein- or enzyme-coding genes. Genes with a gray background represent NFI genes. Dashed lines represent predicted regulatory interactions.

Figure 4. Clustal Omega alignments of PII proteins from P. stutzeri and E. coli. Alignments were constructed using DNA Man and refned manually. Consensus sequences with ≥60% identity are reported below the alignment.

metabolic genes, including the GS-encoding gene glnA (up-regulated 7.61-fold), the nitrogen transcriptional regulation (Ntr) genes, the ammonium transporter gene amtB (up-regulated 55.28-fold), the transcriptional regulator nac (up-regulated 124.98-fold) and the nif-specifc regulator genes nifLA. In fact, we previously demon- strated cross-complementation of the P. stutzeri /or E. coli ntrC (glnG) mutant by the E. coli /or the P. stutzeri ntrC gene34. We have also shown that the purifed E. coli NtrC protein can bind specifcally to the nifLA operon promoter region12. In a sense, this working model supports a similar expression pattern for the NFI genes in E. coli with that observed in P. stutzeri A1501, in that the nif genes were subject to regulation by the NtrC protein via the nif-specifc regulator NifA, suggesting a regulatory coupling of these two diferent evolutionary systems through a direct activating interaction. Te PII protein is recognized as a critical signal transduction protein in bacterial nitrogen metabolism35. E. coli encodes two PII counterparts, GlnB and GlnK, which show 68% and 78% identity, respectively, with the sole A1501 PII protein GlnK and share 67% identity with each other (Fig. 4). Under nitrogen-fxation conditions, E. coli EN-01 glnK mRNA expression was increased by 86.88-fold, while the glnB gene was up-regulated by 2.12-fold. GlnK protein was only detected under nitrogen-fxation conditions (Fig. 2). In A1501, an interaction between GlnK and the C-terminal domain of NifL was observed using the yeast two-hybrid system23, suggesting GlnK-dependent control of NifA activity by NifL. Based on the overall transcription of the nitrogen regulation system in E. coli EN-01, we conjectured that, as reported before35,36, GlnB and GlnK may constitute a two-tiered regulatory system and that their functions may depend on the timing of expression and the levels of accumulation, while glnK may play key roles in nitrogen regulation in E. coli to adjust to perturbations that occur under nitrogen-fxation conditions.

SCieNTiFiC REPOrts | (2018)8:10928 | DOI:10.1038/s41598-018-29204-0 6 www.nature.com/scientificreports/

Rearrangement of the pathways for nitrogen assimilation and alternative nitrogen sources. Glutamate and glutamine provide approximately 75% and 25% of cellular nitrogen, respectively. Nitrogen assimilation must therefore result in the synthesis of these two nitrogen donors37. Te GDH pathway, encoded by the glutamate dehydrogenase gene gdhA, is generally associated with a nitrogen-rich, aerobic environment, whereas the GS-GOGAT pathway, encoded by the GS gene glnA and the glutamate synthase (GOGAT) genes gltBD, is ofen associated with a nitrogen-limited, microaerobic environment37,38. Transcriptomic analyses showed that both operons of the GS-GOGAT system were up-regulated more than 7-fold under nitrogen-fxation conditions, however, the gdhA gene was down-regulated by 2.83-fold (Supplementary Table S3). In accordance with the mRNA results, we also observed that GlnA was up-regulated 7.61-fold and GdhA, encoding GDH, was down-regulated 9.1-fold in the proteomic analysis (Supplementary Table S2). Clearly, the recombinant E. coli EN-01 chooses the high-afnity GS-GOGAT pathway rather than the GDH pathway when environmental nitrogen is depleted. Zimmer et al. previously concluded that E. coli use a range of pathways to scavenge for nitrogen-containing compounds as a frst line of defense against nitrogen starvation39,40. In this work, we observed a rearrangement of transport capacity by EN-01 under nitrogen-fxation conditions, with alterations of 299 transport or membrane protein genes. Notably, the ammonium transporter gene amtB, which forms a transcriptional operon with the PII protein-coding gene glnK, showed a 55-fold enhancement (Fig. 3). Glutamate uptake systems, including the gltIJKL operon, which encodes a glutamate/aspartate uptake system; the glnHPQ operon, which is responsible for a high-afnity glutamine transport system; the gadC gene, which encodes a glutamic acid/γ-aminobutyrate antiporter; and the gltP gene, which encodes a proton/glutamate-aspartate symporter, were up-regulated under nitrogen-fixation conditions. Alternative nitrogen-transport systems, for instance, lysine, arginine, ornithine, histidine, serine, polyamine, peptide, dipeptide and oligopeptide transporter genes, were all signifcantly induced. Additionally, the astABCDE operon, which is responsible for the degradation of arginine into N2-succinyl-L-ornithine, CO2 and ammonia, was also up-regulated (Supplementary Table S3). Nac is a glnK-controlled Ntr protein widely existed in enterobacteria41. Nac is well known to be involved in regulating the expression of multiple pathways to modulate the levels of important cellular metabolic intermedi- ates and thus integrates nitrogen assimilation with other aspects of metabolism41–44. As expected, nac gene expression was drastically up-regulated (over 100-fold) under nitrogen-fxation conditions. gabDTP (γ-aminobutyrate transport and degradation), dppABD (dipeptide transport), and flB (peptidyl-prolyl cis-trans isomerase) gene expression was also increased (Supplementary Table S3). However, codBA (cytosine metabolism), nupC (nucleo- side transport) and ydcSTUVW (putative putrescine transport and degradation) gene expression did not appear to be signifcantly altered. Tese genes were recognized to be controlled by the nac gene in E. coli39. Te unclear inconsistent gene expression obtained in this study may provide important clues into the regulatory disturbances of the NFI island by the nac-conferred nitrogen stress response in recombinant E. coli EN-01.

Metabolic fux shift of E. coli EN-01 under nitrogen-fxation conditions. Biological nitrogen fxation is an energy-dependent process that requires ATP, produced by the catalytic reaction of carbon sources, to break the N≡N bond45. Heterologous NFI expression may place a considerable metabolic burden on the host E. coli. In the present work, EN-01 was observed to shif its metabolic fux to overcome the energy barrier for nitrogen fxation. As the central carbon pathways constitute the backbone of cell metabolism by providing energy, metabolic building blocks, and reducing power for biomass synthesis, we investigated the responses of central carbon metabolism in EN-01. As shown in Fig. 5, a stronger fux toward the pentose phosphate pathway (PPP) was found under nitrogen-fxation conditions. Tus, the PPP was postulated to serve as the primary route for glucose catabolism, since the key genes encoding the rate-limiting enzymes glucose-6-phosphate 1-dehydrogenase (zwf), transketolase (tktAB), transaldolase (talA) and glucose-6-phosphate isomerase (pgi) were highly induced. Te enhanced expression of glucose-6-phosphate 1-dehydrogenase was also detected in proteome analyses (Supplementary Table S2). Conversely, the glyoxylate shunt and Entner-Doudorof pathway were found to be inhibited in EN-01 under nitrogen-fxation conditions. Furthermore, the electron transfer systems commonly involved in energy production46,47 were up-regulated, such as the appABC (encoding cytochrome bd-II oxidase), hyaABCDE (encoding the hydrogenase), flA (encoding the transcriptional activator of formate hydrogenlyase), fdhF and hycABCDEFGH (encoding formate hydrogenlyase complex), ndh (encoding the NADH dehydrogenase), napH (encoding the ferredoxin-type protein), and adhE (encoding the bifunctional acetaldehyde-CoA/ alcohol dehydrogenase) genes. Meanwhile, both the transcription and translation of the formate acetyltransferase 1 gene pfB were enhanced (Supplementary Table S2). Additionally, three alcohol dehydrogenase genes that are reportedly involved in ethanol fermentation were up-regulated. E. coli is an anaerobic fermentative bacterium that produces a variety of organic acids by utilizing glucose under nitrogen-fxation conditions; however, these acids are harmful to nitrogenase activity. In contrast to E. coli, P. stutzeri A1501 is an alkali-producing strain that supplies an optimal environment for nitrogen fxation at pH 7.0–9.0 (Supplementary Fig. S2). Te organic acid content in E. coli wild-type and EN-01, as well as P. stutzeri A1501, were determined by HPLC afer growth under nitrogen-fxation conditions. E. coli wild-type and EN-01 produced a large number of organic acids under nitrogen-fxation conditions (Supplementary Table S4), including pyruvic acid, succinic acid, lactic acid, formic acid and acetic acid, whereas A1501 does not generate these organic acids. Te presence of these organic acids could certainly have negative efects on nitrogenase activity in the recombinant EN-01 strain. Altering the intracellular acid environment may be a feasible method to increase the nitrogen-fxing ability of E. coli EN-01.

Oxygen protection strategy of E. coli EN-01 to cope with O2 toxicity to nitrogenase. Nitrogenase is extremely oxygen sensitive. In fact, purifed nitrogenase, regardless of its source, undergoes extremely rapid 48 and irreversible inactivation by O2 . E. coli EN-01 can grow under both aerobic and anaerobic conditions but only executes nitrogen-fxing activity anaerobically12. Given the importance of oxygen in nitrogen fxation,

SCieNTiFiC REPOrts | (2018)8:10928 | DOI:10.1038/s41598-018-29204-0 7 www.nature.com/scientificreports/

Figure 5. Bioreaction metabolic fux shifs of EN-01 central carbon metabolism under nitrogen-fxation conditions. Changes in the central carbon metabolism network were constructed from microarray data. Arrows indicate the physiological directions of reactions. Red arrows, enhanced expression (P ≤ 0.05 and fold change ≥2); blue arrows, reduced expression (P ≤ 0.05 and fold-change ≥2). Te accumulation of organic acids highlighted in gray was detected in the medium by HPLC. Abbreviations: G6P, glucose 6-phosphate; F6P, fructose 6-phosphate; F1,6P2, fructose 1,6-phosphate; G3P, glyceraldehyde 3-phosphate; 1,3-BPG, 1,3 diphosphoglycerate; 3PG, 3-phosphoglycerate; PEP, phosphoenolpyruvic acid; PYR, pyruvate; AcCoA, acetyl coenzyme A; OAA, oxaloacetate; ICT, isocitrate; AKG, α-ketoglutarate; SUC, succinate; MAL, malate; GLX, glyoxylic acid; 6PG, 6-phosphogluconate; E4P, erythrose 4-phosphate; Ru5P, ribulose-5-phosphate; R5P, ribose-5-phosphate; X5P, xylulose -5-phosphate; S7P, sedoheptulose-7-phosphate; KDPG, 2-keto-3-deoxy-6- phosphogluconic acid; Acetyl-P, acetyl phosphate.

we explored the potential mechanism for oxygen protection in E. coli EN-01 under anaerobic nitrogen-fxing conditions. In the present study, the scavenging capacity of E. coli EN-01 for superoxide radicals, hydrogen peroxide, and hydroxyl radicals appeared to increase under nitrogen-fxation conditions. Te genes sodC and katE, which encode superoxide dismutase and the hydroperoxidase HPII, were up-regulated 2.94-fold and 15.47-fold, respectively, under nitrogen-fxation conditions (Supplementary Table S3). Furthermore, proteomic analysis revealed these two gene products were increased as well (Supplementary Table S2). Other genes such as the hydrogenase 1-coding genes hyaABCDEF, cytochrome bd-II synthetic-related genes, which have higher oxygen afnity than ATP terminal oxidase, and all cytochrome c synthetic genes, were also up-regulated. Te osmC gene, whose product is involved in the detoxifcation of organic hydroperoxides49,50, was also up-regulated 5.81-fold under nitrogen-fxation conditions (Supplementary Table S3). In addition, the yhbO gene was up-regulated 7.5-fold, and its product was only detected under nitrogen-fxation conditions (Supplementary Table S2). YhbO has been shown to function in the response to various stresses. Mutating the yhbO gene in E. coli leads to increased sensi- tivity to heat, oxidative, hyperosmotic, pH and UV stresses51. Although no signifcant changes were detected at the mRNA level, proteomic analysis revealed that Dps was induced 4.06-fold under nitrogen-fxation conditions (Supplementary Table S2). Te Dps protein, a nonspecifc DNA-binding protein, is reported to reduce the production of oxidative radicals through ferroxidase activity and therefore to protect the cell from oxidative stress, UV irradiation, iron and copper toxicity, among others52,53. Taken together, the genes or proteins mentioned above, despite their unclear functions, are thought to be involved in the oxygen stress response under nitrogen-fxation conditions and to simultaneously protect the nitrogenase complex from O2 inhibition by providing a tolerable oxygen environment in E. coli EN-01.

SCieNTiFiC REPOrts | (2018)8:10928 | DOI:10.1038/s41598-018-29204-0 8 www.nature.com/scientificreports/

Discussion To date, the ability to fx nitrogen is exclusively found among bacteria, including green sulfur bacteria, frmibac- teria, actinomycetes, cyanobacteria and all subdivisions of proteobacteria54,55. Genomic analyses have revealed that nitrogen-fxation genes are clustered in most diazotrophs56. Te nif genes were further presumed to have been lost in most bacterial and archaeal lineages during evolution and that horizontal gene transfer played a pivotal role in the recent acquisition of nif genes in some species55–57. In agreement with this hypothesis, several groups have found that transferring nif genes to non-nitrogen fxers successfully yields functional recombinant nitrogen-fxing strains, albeit with far lower nitrogenase activity compared with the donor strains5,11,12,18. Te structural and regulatory systems that control nitrogen fxation in diferent diazotrophs are extremely well conserved, whereas the regulatory mechanisms of nif genes difer from one organism to another28,54. Terefore, for recombinant nitrogen-fxing strains harboring a heterologous gene system, a variety of metabolic penalties and regulatory barriers may exist between the two systems. Understanding how to overcome and escape the metabolic penalties and regulatory barriers is essential to fne-tune acquired nitrogen-fxing capacity of the non-diazotrophic host. In this work, transcriptomic and proteomic analysis separately identifed 1156 genes and 138 proteins whose expression was signifcantly altered when E. coli EN-01 was incubated under nitrogen-fxation conditions. Our results highlight the metabolic penalties and regulatory barriers for nitrogen signal transduction and metabolism brought by excess fxed nitrogen in the medium. Tese genes and proteins, particularly the 24 genes that over- lapped between the two transcriptomic analyses (the previous transcriptomic analysis of P. stutzeri A1501 and the E. coli analysis presented in this work), are potential targets for subsequent modifcation in E. coli EN-01 to optimize nitrogen-fxing capacity. Te nitrogen regulatory cascade of P. stutzeri A1501 comprises the AmtB–GlnK–NtrBC general nitrogen regulation proteins and the nif-specifc regulatory protein NifLA22,24,26,31. E. coli has a similar nitrogen regulatory system to P. stutzeri but harbors an additional PII protein and a cascade regulation by the Nac regulatory pro- tein32,33. In the present work, a similar expression network of NFI genes in E. coli to that operating in P. stutzeri A1501 was predicted based on the expression patterns of rpoN, glnD, glnA, glnK, ntrBC, nifLA and other nitrogen regulation-related genes (Fig. 3). Given that the purifed E. coli NtrC protein can bind specifcally to the A1501 nifLA operon promoter region12, we postulated the regulatory coupling of these two evolutionary divergent systems through a direct activating interaction. Additionally, the nac gene was also predicted to be involved in nif gene regulation in EN-01 although the mechanism requires further investigation. Interestingly, an expression diference was observed in the two PII counterparts GlnB and GlnK in E. coli. In principle, the functions of GlnK and GlnB overlap in Enterobacteriaceae. However, both proteins exhibited distinct physiological roles in the regulation of nitrogen assimilation over a wide range of environmental conditions, despite having distinct expression patterns. Te glnB promoter contains a putative -10 sequence for RpoS activation, while the transcription of glnK is induced by NtrC and RpoN35. We therefore presumed that glnK may play a more important role than GlnB in nitrogen regulation in E. coli under nitrogen-fxation conditions. Notably, the PST1302-1306 cluster within the NFI showed unregulated expression in E. coli EN-01 but was highly up-regulated in P. stutzeri A1501 under nitrogen-fxation conditions24. Te weak transcription of these nitrogenase component genes may predict imper- fections and the necessity for further rebuilding of nif gene expression regulation in E. coli EN-01. A core challenge for diazotrophs is the need to generate sufcient energy to drive nitrogen fxation. In the present work a strong shif in the metabolic fux of the EN-01 strain was observed under nitrogen-fxation conditions. Te key enzyme of the PPP was up-regulated, but the glyoxylate shunt and Entner-Doudorof pathway were inhibited. Tus, the PPP pathway was concluded to serve as the primary route for glucose catabolism to support nitrogen fxation. Moreover, anaerobic respiration-related electron transfer systems and substances were induced. Tese data highlight the energy acquisition pathway of EN-01 to adapt to nitrogen fxation. However, as an anaerobic fermentative bacterium, EN-01 produces a large number of organic acids in the media, which may negatively afect nitrogenase activity, since the donor strain P. stutzeri A1501 is an alkali strain. Changing the intracellular acid environment may thus be a feasible method to increase the nitrogen-fxing ability of E. coli EN-01. We identifed a set of genes that might be involved in protecting nitrogenase from oxygen via scavenging of superoxide radicals, hydrogen peroxide, and hydroxyl radicals. Two proteins, YhbO and Dps, also appeared to be involved in the oxygen protection of nitrogenase by protecting the cell from oxidative stress. Tese genes or proteins may protect the nitrogenase complex from O2 inhibition by establishing a tolerable oxygen environment in E. coli EN-01. Terefore, we speculate that improving the expression of antioxidant capacity-related pathway genes may satisfy the need for oxygen protection under microaerobic conditions and perhaps drive the metabolic fux shif from anaerobic respiration to aerobic respiration to produce more energy and reduce the energy requirements of the nitrogen fxation. Te nitrogenase activity of E. coli EN-01 was much lower than that in P. stutzeri A1501. According to the data in this work, this diference may be due to the regulatory divergence of nitrogen assimilation, the inactivation of the gene operon PST1302-1306, and the production of organic acids. Another factor may also be involved. In our recent work, a species-specifc regulatory ncRNA nfS identifed in P. stutzeri A1501, could directly base paired with the mRNA of the nitrogenase component nifK to enhance translation efciency and transcript half-life, thereby regulating nitrogenase biosynthesis27. Blast results showed that no sequence homologous to A1501 nfS was present in the E. coli genome. Introduction of nfS regulation into E. coli EN-01 may be a useful way to refne NFI gene regulation. From an agri-biotechnology perspective, the successful engineering of an N2-fxing organism would significantly reduce the need for nitrogen fertilizers15,16,54. Biosynthetic techniques have been used by several groups to improve or optimize nitrogenase activity in metabolically engineered strains. Wang X. et al. replaced the nif regulatory elements in the recombinant E. coli strain with a T7 RNA polymerase–LacI expression system to overcome regulatory barriers, and the newly constructed T7-dependent nif system bypassed the original complex

SCieNTiFiC REPOrts | (2018)8:10928 | DOI:10.1038/s41598-018-29204-0 9 www.nature.com/scientificreports/

regulatory circuits with minor physiological limitations9. Subsequent work from the same group pinpointed the electron-transfer components from plant organelles that can be used to support nitrogenase activity and reduction of the number of target genes required to engineer nitrogen fxation in plants58. Another group observed that fve separate gene clusters, as well as a combined cluster from Paenibacillus sp. WLY78 and Klebsiella oxytoca, increase the nitrogenase activity of the recombinant E. coli 78-7, which harbors a minimal nif gene cluster from Paenibacillus sp. WLY7811,19. In summary, the results presented in this work demonstrate the overall expression pattern of E. coli EN-01 genes under nitrogen fxation and revealed the regulatory coupling of host genes and the heterologous NFI that supports functional nitrogenase activity in the engineered E. coli strain. We also uncovered the metabolic penalties and regulatory barriers that lead to the reduced nitrogenase activity in E. coli EN-01 and provided clues for biotechnological applications to generate new hypotheses concerning gene regulation and fux, which may lead to an improved ability to fx nitrogen in E. coli. Tis work may shed light on the metabolic penalties and coordinated patterns between host and horizontally transferred genes and provide the knowledge necessary to pave the way for engineered nitrogen fxation. Methods Bacterial strains and growth conditions. Te strains and plasmids used in this study are described in Table S5. The recombinant strain E. coli EN-01 bearing an NFI was grown at 37 °C in LB medium12. Chloramphenicol, tetracycline and hygromycin were added to media at a concentration of 20 μg/mL, 10 μg/mL and 100 μg/mL, respectively, as required. P. stutzeri A1501 was cultured at 30 °C in LB medium or minimal lac- tate-containing medium (medium K) as described previously20.

Nitrogenase activity assays. To measure nitrogenase activity in E. coli EN-01, cells were cultured overnight to an OD600 of 2.0 at 37 °C in LB, and 0.2 mL cultures were directly added to 6 mL NFDM minimal medium −1 (20 g glucose, 0.7 g MgSO4·7H2O, 25 mg Na2MoO4·H2O, 3.6 g ferric citrate, 68 g KH2PO4, 241 g K2HPO4, L , pH 7.4) without nitrogen in 10 mL serum bottles. Tese cultures were sealed and then incubated at 30 °C for 20 h with constant shaking (200 rpm), afer which 10% acetylene without oxygen was injected to determine ethylene production by gas chromatography, as described by Cannon et al.59. Te nitrogenase activity of P. stutzeri was determined in N-free minimal medium at an OD600 of 0.1 at 30 °C under an argon atmosphere containing 0.5% oxygen and 10% acetylene, according to the protocol described by Desnoues et al.60. Nitrogenase-specifc activity is expressed as nmol ethylene per mg protein per h. Each experiment was repeated at least three times.

Total RNA preparation. E. coli EN-01 cells were cultured overnight to an OD600 of 2.0 at 37 °C in LB, and 0.2 mL cultures were directly added to 6 mL NFDM medium containing 20 mM ammonium (nitrogen-repression conditions) or without ammonium (nitrogen-fxation conditions) in 10 mL serum bottles, respectively. Tese cultures were sealed and incubated at 30 °C for 20 h with constant shaking (200 rpm). Te cell densities of EN-01 under anaerobic nitrogen-fxation or nitrogen-repression conditions afer 20 h reached 3.43 × 108 CFU/mL and 5.42 × 108 CFU/mL, respectively, indicating similar growth status. Te EN-01 cells were then centrifuged at 12,000 g for 5 min. Subsequently, cell pellets were rapidly frozen in liquid nitrogen and stored at −80 °C. Total RNA was isolated using TRIzol Reagent (Invitrogen, Carlsbad, CA). To avoid possible DNA contamination, an additional DNase I digestion was performed at 37 °C for 30 min, and the samples were chilled on ice. Total RNA was purifed using RNeasy columns, resuspended in RNase-free water and quantitated on a NanoVue plus.

Microarray analysis. For microarrays, standard methods were used for cDNA synthesis, fragmentation, and terminal biotin labeling based on Afymetrix protocols. Labeled cDNA was hybridized to the Afymetrix E. coli Genome 2.0 array. Hybridized arrays were stained with streptavidin-phycoerythrin using an Afymetrix Fluidic Station. Afer staining, arrays were scanned with an Afymetrix GeneChip Scanner 3000 based on the total signal intensity. Te resulting microarray data were analyzed using Afymetrix sofware (MAS 5.0). Consensus “detec- tion p-values”, “change p-values”, and “mean expression ratios” were calculated. All signal intensities with mean expression ratios above 2 were considered signifcant changes if the p-value was below 0.05. Complete microarray data have been deposited in the Gene Expression Omnibus (GEO) database under accession number GSE37780.

Quantitative real-time PCR. To investigate the expression of NFI genes in E. coli EN-01, real-time quantitative PCR (qPCR) was performed using an ABI 7500 Real-Time PCR System. Primers pairs used for qPCR were designed using DNAman and are listed in Table S6. RNA was extracted from cells treated under the same conditions described for the microarray analyses. Reverse transcription was performed using a cDNA synthesis kit (Promega, USA). SYBR Green premix was used to detect PCR amplifcation. Te 16S RNA gene was used to normalize the results. Additionally, qPCR was also used to verify microarray results.

Proteome analysis by two-dimensional PAGE. As previously described, E. coli EN-01 cells were cultured under the anaerobic nitrogen-fxation or nitrogen-repression conditions. Te cells were harvested via high-speed centrifugation (12,000 g, 5 min and 4 °C). Cell pellets were washed three times with PBS bufer and suspended in lysis bufer (8 M urea, 2 M thiourea, 0.5% w/v CHAPS, 2% v/v carrier ampholyte, 1% w/v DTT and 1 mM PMSF) for sonication on ice. Te samples were centrifuged at 14000 g, 4 °C for 30 min, then, the supernatant was taken for the Bradford Protein assay and frozen at −80 °C. Te details of two-dimensional protein gel electrophoresis and 2-DE analysis were performed according to the protocol described by Zhengfu Zhou et al.61. Briefy, an equal amount of total protein extract (1000 µg) was isoe- lectrically separated on a 17 cm immobilized pH gradient strip (Bio-Rad, USA) with a linear pH gradient from 4 to 7. Separation along the second-dimension was performed using vertical 12% gels, followed by Coomassie Brilliant Blue G-250 staining visualization and image acquisition. Gels were scanned with a PowerLook 1000

SCieNTiFiC REPOrts | (2018)8:10928 | DOI:10.1038/s41598-018-29204-0 10 www.nature.com/scientificreports/

(UMAX Technologies Inc., Dallas, TX) and the gel images were analyzed with PDQuest V7.3.0 (Bio-Rad Laboratories) sofware according to the manufacture’s protocol. Te protein spots with intensity levels greater than 2.0 or less than 0.5 were picked and subjected to mass spectrometric (MS) analysis. For MALDI-TOF MS analysis, the protein spots with signifcant diference were excised from the gels for tryptic digestion as described61. Te peptides were analyzed on a 4700 Proteomics Analyzer (Applied Biosystems, USA). Proteins were identifed using GPS Explorer sofware V3.5 (Applied Biosystems) and the function was quoted from the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov) database.

Measurement of intercellular organic acid content. E. coli DH10B and its recombinant strain EN-01 were cultured under anaerobic nitrogen-fxation conditions, and cells were collected at 0 h and 45 h. A1501 was grown under microaerobic nitrogen-fxation conditions, and the bacteria were centrifuged at 0 h and 10 h. Te cell suspension was centrifuged at 12000 g, 4 °C for 10 min, and the supernatant was used to analyze the organic acid content. Organic acid concentrations were determined using an HPLC (1200 series, Agilent Technologies) equipped with a Bio-Rad HPX-87H column with 5 mM H2SO4 as the mobile phase (0.6 ml/min, 50 °C). References 1. Ochman, H., Lawrence, J. G. & Groisman, E. A. Lateral gene transfer and the nature of bacterial innovation. Nature 405, 299–304, https://doi.org/10.1038/35012500 (2000). 2. Pierneef, R., Cronje, L., Bezuidt, O. & Reva, O. N. Pre_GI: a global map of ontological links between horizontally transferred genomic islands in bacterial and archaeal genomes. Database bav 58, https://doi.org/10.1093/database/bav058 (2015). 3. Melnyk, R. A. & Coates, J. D. Te Perchlorate Reduction Genomic Island: Mechanisms and Pathways of Evolution by Horizontal Gene Transfer. BMC Genomics 16, 862, https://doi.org/10.1186/s12864-015-2011-5 (2015). 4. Dixon, R. & Postgate, J. Transfer of nitrogen-fxation genes by conjugation in Klebsiella pneumoniae. Nature 234, 47–48, https://doi. org/10.1038/234047a0 (1971). 5. Dixon, R. & Postgate, J. Genetic transfer of nitrogen fxation from Klebsiella pneumoniae to Escherichia coli. Nature 237, 102–103, https://doi.org/10.1038/237102a0 (1972). 6. Streicher, S., Gurney, E. & Valentine, R. C. Transduction of the nitrogen fxation genes in Klebsiella pneumoniae. Proc Natl Acad Sci USA 68, 1174–1177 (1971). 7. Postgate, J. & Kent, H. M. Qualitative Evidence for Expression of Klebsiella pneumoniae nif in Pseudomonas putida. Journal of General Microbiology 133, 2563–2566, https://doi.org/10.1099/00221287-133-9-2563 (1987). 8. Temme, K., Zhao, D. & Voigt, C. A. Refactoring the nitrogen fxation gene cluster from Klebsiella oxytoca. Proc Natl Acad Sci USA 109, 7085–7090, https://doi.org/10.1073/pnas.1120788109 (2012). 9. Wang, X. et al. Using Synthetic Biology to Distinguish and Overcome Regulatory and Functional Barriers Related to Nitrogen Fixation. Plos One 8, e68677, https://doi.org/10.1371/journal.pone.0068677 (2013). 10. Setten, L. et al. Engineering Pseudomonas protegens Pf-5 for Nitrogen Fixation and its Application to Improve Plant Growth under Nitrogen-Defcient Conditions. Plos One 8, e63666, https://doi.org/10.1371/journal.pone.0063666 (2013). 11. Wang, L. et al. A Minimal Nitrogen Fixation Gene Cluster from Paenibacillus sp. WLY78 Enables Expression of Active Nitrogenase in Escherichia coli. Plos Genetics 9, e1003865, https://doi.org/10.1371/journal.pgen.1003865 (2013). 12. Han, Y. L. et al. Interspecies Transfer and Regulation of Pseudomonas stutzeri A1501 Nitrogen Fixation Island in Escherichia coli. J. Microbiol. Biotechnol 25, 1339–1348, https://doi.org/10.4014/jmb.1502.02027 (2015). 13. Cheng, Q., Day, A., Dowson-Day, M., Shen, G. & Dixon, R. The Klebsiella pneumoniae nitrogenase Fe proteingene (nifH) functionally substitute for the chlL gene in Chlamydomonas reinhardtii. Biochemical and Biophysical Research Communication 329, 966–975, https://doi.org/10.1016/j.bbrc.2005.02.064 (2005). 14. Cheng, Q. Perspectives in Biological Nitrogen Fixation Research. Journal of Integrative Plant Biology 50, 784–796, https://doi. org/10.1111/j.1744-7909.2008.00700.x (2008). 15. Geddes, B. A. et al. Use of plant colonizing bacteria as chassis for transfer of N2-fxation to cereals. Current opinion in biotechnology 32, 216–222, https://doi.org/10.1016/j.copbio.2015.01.004 (2015). 16. Vicente, E. J. & Dean, D. R. Keeping the nitrogen-fixation dream alive. Proc Natl Acad Sci USA 114, 3009–3011, https://doi. org/10.1073/pnas.1701560114 (2017). 17. Mus, F. et al. Symbiotic nitrogen fxation and the challenges to its extension to nonlegumes. Applied and environmental microbiology 82, 3698–3710, https://doi.org/10.1128/AEM.01055-16 (2016). 18. Yang, J., Xie, X., Wang, X., Dixon, R. & Wang, Y. Reconstruction and minimal gene requirements for the alternative iron-only nitrogenase in Escherichia coli. Proc Natl Acad Sci USA 111, E3718–E3725, https://doi.org/10.1073/pnas.1411185111 (2014). 19. Li, X., Liu, Q., Liu, X., Shi, H. & Chen, S. Using synthetic biology to increase nitrogenase activity. Microbial Cell Factories 15, 43, https://doi.org/10.1186/s12934-016-0442-6 (2016). 20. Lin, M. & You, C. Denitrifcation and nitrogen fxation by Alcaligenes faecalis. Acta Agriculturae Nucleatae Sinica 1, 1–7 (1987). 21. Rediers, H. et al. Development and application of a dapB-based in vivo expression technology system to study colonization of rice by the endophytic nitrogen-fixing bacterium Pseudomonas stutzeri A15. Appl Environ Microbiol 69, 64–6874, https://doi. org/10.1128/AEM.69.11.6864-6874 (2003). 22. Xie, Z. H. et al. Interaction between NifL and NifA in the nitrogen-fxing Pseudomonas stutzeri A1501. Microbiology 152, 3535–3542, https://doi.org/10.1099/mic.0.29171-0 (2006). 23. He, S. et al. Involvement of GlnK, a PII protein, in control of nitrogen fxation and ammonia assimilation in Pseudomonas stutzeri A1501. Arch Microbiol 190, 1–10, https://doi.org/10.1007/s00203-008-0354-x (2008). 24. Yan, Y. et al. Global transcriptional analysis of nitrogen fxation and ammonium repression in root-associated Pseudomonas stutzeri A1501. BMC Genomics 11, 11, https://doi.org/10.1186/1471-2164-11-11 (2011). 25. Yan, Y. et al. Nitrogen fxation island and rhizosphere competence traits in the genome of root-associated Pseudomonas stutzeri A1501. Proc Natl Acad Sci USA 105, 7564–7569, https://doi.org/10.1073/pnas.0801093105 (2008). 26. Zhang, T. et al. Involvement of the ammonium transporter AmtB in nitrogenase regulation and ammonium excretion in Pseudomonas stutzeri A1501. Research in Microbiology 163, 332–339, https://doi.org/10.1016/j.resmic.2012.05.002 (2012). 27. Zhan, Y. et al. Te novel regulatory ncRNA, NfS, optimizes nitrogen fxation via base pairing with the nitrogenase gene nifK mRNA in Pseudomonas stutzeri A1501. Proc Natl Acad Sci USA 113, E4348–E4356, https://doi.org/10.1073/pnas.1604514113 (2016). 28. Dixon, R. & Kahn, D. Genetic regulation of biological nitrogen fxation. Nat. Rev. Microbiol 2, 621–631, https://doi.org/10.1038/ nrmicro954 (2004). 29. Imperial, J., Ugalde, R. A., Shah, V. K. & Brill, W. J. Role of the nifQ Gene Product in the Incorporation of Molybdenum into Nitrogenase in Klebsiella pneumonia. Journal of Bacteriology 158, 187–194 (1984). 30. Roberts, G. P., MacNeil, T., MacNeil, D. & Brill, W. J. Regulation and Characterization of Protein Products Coded by the nif (Nitrogen Fixation) Genes of Klebsiella pneumoniae. Journal of Bacteriology 136, 267–279 (1978).

SCieNTiFiC REPOrts | (2018)8:10928 | DOI:10.1038/s41598-018-29204-0 11 www.nature.com/scientificreports/

31. Lin, M., Yan, Y., Lu, W., Zhan, Y. & Zhang, Y. Regulatory Coupling of the Nitrogen and Carbon Metabolism in Nitrogen-Fixing Pseudomonas stutzeri A1501. Biological Nitrogen Fixation (ed. De Bruijn FransJ.) 1, 109–119 (Wiley-Blackwell, Hoboken, NJ, 2015). 32. Van Heeswijk, W. C. et al. An alternative PII protein in the regulation of glutamine synthetase in Escherichia coli. Molecular Microbiology 21, 133–146, https://doi.org/10.1046/j.1365-2958.1996.6281349.x (1996). 33. Muse, W. B. & Bender, R. A. Te nac (Nitrogen Assimilation Control) Gene from Escherichia coli. Journal of Bacteriology 180, 1166–1173 (1998). 34. Ke, Q., Chen, Q., Zhan, Y., Lu, W. & Yan, Y. Functional Complement Between General Nitrogen Regulator GlnG in Escherichia coli and General Nitrogen Regulator NtrC in Pseudomonas. Biotechnology Bulletin 32, 190–197, https://doi.org/10.13560/j.cnki.biotech. bull.1985.2016.04.025 (2016). 35. Arcondéguy, T., Jack, R. & Merrick, M. PII Signal Transduction Proteins, Pivotal Players in Microbial Nitrogen Control. Microbiology and Molecular Biology Reviews 65, 80–105, https://doi.org/10.1128/MMBR.65.1.80-105.2001 (2001). 36. Atkinson, M. R., Blauwkamp, T. A. & Ninfa, A. J. Context-Dependent Functions of the PII and GlnK Signal Transduction Proteins in Escherichia coli. Journal of Bacteriology 184, 5364–5375, https://doi.org/10.1128/JB.184.19.5364-5375.2002 (2002). 37. Reitzer, L. & Schneider, B. L. Metabolic Context and Possible Physiological Temes of δ54-Dependent Genes in Escherichia coli. Microbiology and Molecular Biology Reviews 65, 422–444 (2001). 38. Reitzer, L. Nitrogen Assimilation and Global Regulation in Escherichia coli. Annual Review of Microbiology 57, 155–176, https://doi. org/10.1146/annurev.micro.57.030502.090820 (2003). 39. Zimmer, D. P. et al. Nitrogen regulatory protein C-controlled genes of Escherichia coli: Scavenging as a defense against nitrogen limitation. Proc Natl Acad Sci USA 97, 14674–14679, https://doi.org/10.1073/pnas.97.26.14674 (2000). 40. Wang, J., Yan, D., Dixon, R. & Wang, Y. Deciphering the Principles of Bacterial Nitrogen Dietary Preferences: a Strategy for Nutrient Containment. mBio 7, e00792–16, https://doi.org/10.1128/mBio.00792-16 (2016). 41. Muse, W. B. & Bender, R. A. The nac (nitrogen assimilation control) gene from Escherichia coli. Journal of Bacteriology 180, 1166–1173 (1998). 42. Blauwkamp, T. A. & Ninfa, A. J. Physiological role of the GlnK signal transduction protein of Escherichia coli: survival of nitrogen starvation. Molecular Microbiology 46, 203–214, https://doi.org/10.1046/j.1365-2958.2002.03153.x (2002). 43. Goss, T. J., Perez-Matos, A. & Bender, R. A. Roles of glutamate synthase, gltBD, and gltF in nitrogen metabolism of Escherichia coli and Klebsiella aerogenes. Journal of Bacteriology 183, 6607–6619, https://doi.org/10.1128/JB.183.22.6607-6619.2001 (2001). 44. Poggio, S., Domeinzain, C., Osorio, A. & Camarena, L. Te nitrogen assimilation control (Nac) protein represses asnC and asnA transcription in. Escherichia coli. FEMS Microbiol. Lett. 206, 151–156, https://doi.org/10.1111/j.1574-6968.2002.tb11001.x (2002). 45. Canfeld, D. E., Glazer, A. N. & Falkowski, P. G. Te Evolution and Future of Earth’s Nitrogen Cycle. Science 330(192–196), 330: 192–196 (2010). 46. Borisova, V. B., Gennis, R. B., Hemp, J. & Verkhovsky, M. I. Te cytochrome bd respiratory oxygen reductases. Biochimica et Biophysica Acta 1807, 1398–1413, https://doi.org/10.1016/j.bbabio.2011.06.016 (2011). 47. Wissenbach, U., Kröger, A. & Unden, G. Te specifc functions of menaquinone and demethylmenaquinone in anaerobic respiration with fumarate, dimethylsulfoxide, trimethylamic N-oxide and nitrate by Escherichia coli. Archives of Microbiology 154, 60–66, https://doi.org/10.1007/BF00249179 (1990). 48. Gallon, J. R. Reconciling the incompatible: N2 fxation and O2. New Phytologist 122, 571–609 (1992). 49. Mongkolsuk, S., Praituan, W., Loprasert, S., Fuangthong, M. & Chamnongpol, S. Identifcation and characterization of a new organic hydroperoxide resistance (ohr) gene with a novel pattern of oxidative stress regulation from Xanthomonas campestris pv. phaseoli. Journal of Bacteriology 180, 2636–2643 (1998). 50. Atichartpongkul, S. et al. Bacterial Ohr and OsmC paralogues defne two protein families with distinct functions and patterns of expression. Microbiology 147, 1775–1782, https://doi.org/10.1099/00221287-147-7-1775 (2001). 51. Abdallah, J., Caldas, T., Kthiri, F., Kern, R. & Richarme, G. YhbO Protects Cells against Multiple Stresses. American Society for Microbiology 189, 9140–9144, https://doi.org/10.1128/JB.01208-07 (2007). 52. Buda, F. B., Ensing, B. E., Gribnau, M. C. M. & Baerends, E. J. O2 evolution in the Fenton reaction. Chemistry 9, 3436–3444, https:// doi.org/10.1002/chem.200204444 (2003). 53. Nair, S. & Finkel, S. E. Dps Protects Cells against Multiple Stresses during Stationary Phase. Journal of Bacteriology 186, 4192–4198 (2004). 54. Claudine, E. Associative and endophytic nitrogen-fxing bacteria and cyanobacterial associations (ed. Claudine, E. & Newton, W. E.) 1–16, Vol. 5. Dordrecht: Springer (2007). 55. Raymond, J., Siefert, J. L., Staples, C. R. & Blankenship, R. E. Te Natural History of Nitrogen Fixation. Molecular Biology & Evolution 21, 541–554, https://doi.org/10.1093/molbev/msh047 (2004). 56. Venieraki, A. et al. Te nitrogen-fxation island insertion site is conserved in diazotrophic Pseudomonas stutzeri and Pseudomonas sp. isolated from distal and close geographical regions. Plos One 9, e105837, https://doi.org/10.1371/journal.pone.0105837 (2014). 57. Xiong, J., Fischer, W. M., Inoue, K., Nakahara, M. & Bauer, C. E. Molecular evidence for the early evolution of photosynthesis. Science 289, 1724–1730, https://doi.org/10.1126/science.289.5485.1724 (2000). 58. Yang, J., Xie, X., Yang, M., Dixon, R. & Wang, Y. Modular electron-transport chains from eukaryotic organelles function to support nitrogenase activity. Proceedings of the National Academy of Sciences of the United States of America 114, 2460–2465, https://doi. org/10.1073/pnas.1620058114 (2017). 59. Cannon, F. C., Dixon, R., Postgate, J. R. & Primrose, S. B. Chromosomal integration of Klebsiella nitrogen fixation genes in Escherichia coli. Journal of General Microbiology 80, 227–239, https://doi.org/10.1099/00221287-80-1-227 (1974). 60. Desnoues, N. et al. Nitrogen fxation genetics and regulation in a Pseudomonas stutzeri strain associated with rice. Microbiology 149, 2251–2262, https://doi.org/10.1099/mic.0.26270-0 (2003). 61. Zhou, Z. F. et al. Genome-wide transcriptome and proteome analysis of Escherichia coli expressing IrrE, a global regulator of Deinococcus radiodurans. Molecular Biosystems 7, 1613–1620, https://doi.org/10.1039/c0mb00336k (2011). Acknowledgements Tis work was supported by National Natural Science Foundation of China (31230004, 31470205, 31470174 and 31770067), the National Basic Research Program of China (2015CB755700), and Ministry of Agriculture (Transgenic Program, No. 2016ZX08009003-002). This study was also conducted with the support of the Agricultural Science and Technology Innovation Program (2014-2017), Fundamental Research Funds for Central Non-proft Scientifc Institution (0392017002) and the Guangdong Innovative and Entrepreneurial Research Team Program (No. 2013S033). We would like to thank Dr. Claudine Elmerich and Dr. Ray Dixon for many helpful discussions.

SCieNTiFiC REPOrts | (2018)8:10928 | DOI:10.1038/s41598-018-29204-0 12 www.nature.com/scientificreports/

Author Contributions Y.Y. and M.L. conceived and designed the experiments. Z.Y., Y.H., Y.M., Q.C. and Y.Z. performed the experiments. Z.Y., Y.Y. and M.L. contributed to the writing of the manuscript. Z.Y., W.L., L.C., M.H., Y.Y. and S.C. contributed to the collection and analysis of data. All authors (Z.Y., Y.H., Y.M., Q.C., Y.Z., W.L., L.C., M.H., S.C., Y.Y. & M.L.) discussed the results, commented and approved the version of fnal manuscript. Additional Information Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-018-29204-0. Competing Interests: Te authors declare no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. Te images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

SCieNTiFiC REPOrts | (2018)8:10928 | DOI:10.1038/s41598-018-29204-0 13